|It is important not to burden cellular automata with any expectation that they will solve "real world" problems. Rather, they are a tool for understanding how the world works at a fundamental level.|
In 1983, while I was getting my introduction to cellular automata and to a certain class of graphical programming courtesy Fredkin's automata mentioned above, Stephen Wolfram was classifying a simple set of 1D cellular automata into Class 1 (stable/repetitive), Class 2 (nested), Class 3 (random) and Class 4.
In the same era others, typified by the Santa Fe Institute, came to that oft elusive boundary region between order and chaos from other directions, labeled, amongst other things, as "complex systems".
Wolfram published A New Kind of Science (NKS) last year, building of that early work, and I am more than disappointed that I am not going to be able to attend the first NKS conference in Boston the weekend after next.
Conway's Game of Life is by a big margin the most studied Class 4 system.
I am trying to put the finishing touches to an article which attempts to account for some of the problem in isolating Class 4 behaviour, in part by noting that there really isn't such a thing as a distinctive Class 4 outcome corresponding to those which obviously exist for Classes 1-3, so I should not try to preempt that here.
More and more, the relationship of Life in a Tube to standard Life seems to represent a significant characterisitic of Class 4 systems ... by constraining them with a small amount of additional data, such as the tube circumference, they can be made to exhibit a substantial additional range of unanticipated behaviour.
Biology also typifies the basic idea of complex systems as the emergent behaviour of a large population of simple systems locally interacting in parallel. This is now the way I look at the universe at large, the laws of physics appearing to be the same everywhere.
Cellular automata are a simpler toy/test bed than even biology for exploring the universals of complex systems.